Hi Lukas by the way I found this very strange, however you are right, it seems that pacemaker left old haproxy instances alive (I moved the service between the nodes to simulate a shutdown on one node) - I didn’t noticed this, as it should kill every instance on one node before starting them on another node.
after moving the resource to another node, killing the survived instances and moving the service back to the node everything seems working properly. sorry for the wrong report cheers Marco > On 09 Sep 2016, at 12:31, Lukas Tribus <[email protected]> wrote: > > Ciao Marco, > > > I assume the old process did not get the signal and continues to serve > requests with the old configuration. > > Can you confirm the number of haproxy processes running is more than you > expect? Are you using nbproc or single process mode (the latter is the > default)? > > > Does the PID file contain the correct PID? > > > If the old haproxy instance is not getting the signal it will continue to > serve requests and the kernel will load-balance between the old and the new > instance, leading to the behavior you are describing. > > > > We really need an no-reuseport knob to confirm those kinds of issues... > > > Regarding the actual problem, I would suggest to upgrade to latest stable > release 1.6.9 first of all. Then we can actually start troubleshooting, but > there are important bugfixes in those 4 releases. > > > > cheers, > > lukas > >
